Date: Fri, 2 Jun 1995 11:24:45 -0700 (PDT) From: "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com> To: henrich@crh.cl.msu.edu (Charles Henrich) Cc: freebsd-hackers@freebsd.org Subject: Re: A performance mystery Message-ID: <199506021824.LAA07817@gndrsh.aac.dev.com> In-Reply-To: <199506021525.IAA11663@freefall.cdrom.com> from "Charles Henrich" at Jun 2, 95 11:24:55 am
next in thread | previous in thread | raw e-mail | index | archive | help
> > > Try running ``iozone 128 8192'' to make sure you get well outside of > > the buffer cache. I remeber you had a lot of memory in this box > > so your iozone read results where most likely pure cache unless you > > ran something other than auto. Better yet run it on the raw disk to > > see what the drive can really deliver. > > Iozone was run with "200" to work on 200mb files, to ensure the memory had no > real effect. Both systems we're freshly formatted about an hour before the > test using the 2.0.5A installer. With the iozone numbers being so equivlant > (Seacrate 2.3mb write/4.2 read, Connor 2.2write/4.3 read [all within test > noise]) its probably not the disk. Okay, try a memory bandwidth benchmark to see if we have something there that is grossly different. A quick and easy bcopy test is to run iozone that does not remove the temp file on a 1/4 memory size file repeatedly, that should hit the buffer cache totatally and give us ~ bcopy rates. -- Rod Grimes rgrimes@gndrsh.aac.dev.com Accurate Automation Company Custom computers for FreeBSD
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199506021824.LAA07817>